Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implemented method of processing voice utterances, the method comprising: receiving a plurality of user inputs directed at a device topology user interface used to update a device topology representation, the device topology representation defining a plurality of smart devices associated with one or more structures and a location of each of the plurality of smart devices within the one or more structures; adding, to the device topology representation and based on the plurality of user inputs directed at the device topology user interface: a given smart device, a device identifier of the given smart device, the one or more structures, one or more rooms as a given location of the given smart device within the one or more structures, and one or more device attributes for the given smart device, wherein the one or more device attributes for the given smart device are associated with one or more action the given smart device is configured to perform; subsequent to adding to the device topology representation based on the plurality of user inputs directed to the device topology representation: receiving, by one or more computing devices, data indicative of a voice utterance, the voice utterance provided by a user to a control device; accessing, by the one or more computing devices, the device topology representation; determining, by the one or more computing devices, a transcription of the voice utterance based on the device topology representation, wherein determining the transcription of the voice utterance based on the device topology representation comprises: identifying one or more biasing terms or phrases associated with the accessed device topology representation, biasing a general language model based on one or more of the identified biasing terms or phrases associated with the accessed device topology representation, and determining the transcription based on the biasing; and determining, by the one or more computing devices, one or more selected devices, from the plurality of smart devices, and one or more corresponding actions to be performed by the one or more selected devices based at least in part on the determined transcription and the device topology representation.
2. The computer-implemented method of claim 1 , wherein determining, by the one or more computing devices, the transcription of the voice utterance further comprises: providing, by the one or more computing devices, data indicative of the voice utterance to the general language model; determining, by the one or more computing devices, a first transcription of the voice utterance based at least in part on the general language model; and determining, by the one or more computing devices, a first confidence score associated with the first transcription.
This invention relates to voice transcription systems that improve accuracy by leveraging both general and domain-specific language models. The problem addressed is the inherent limitations of general-purpose language models when transcribing specialized or domain-specific speech, which often results in lower accuracy and higher error rates. The method involves processing a voice utterance using a general language model to generate an initial transcription and associated confidence score. This transcription serves as a baseline, which can then be refined or validated using additional domain-specific models or techniques. The confidence score helps assess the reliability of the transcription, enabling further processing steps such as error correction, model selection, or user feedback integration. By combining the general model's broad knowledge with domain-specific refinements, the system aims to achieve higher transcription accuracy across diverse use cases. The approach is particularly useful in applications where speech varies significantly from standard language patterns, such as technical jargon, medical terminology, or industry-specific vocabulary. The method ensures that the transcription process is adaptive and context-aware, improving usability in specialized domains.
3. The computer-implemented method of claim 2 , wherein determining, by the one or more computing devices, the transcription of the voice utterance further comprises: providing, by the one or more computing devices, the data indicative of the voice utterance to a personalized language model, the personalized language model being determined based at least in part on the device topology representation; determining, by the one or more computing devices, a second transcription of the voice utterance based at least in part on the personalized language model; determining, by the one or more computing devices, a second confidence score associated with the second transcription; and selecting, by the one or more computing devices either the first transcription or the second transcription to be interpreted based at least in part on the first and second confidence scores.
This invention relates to voice recognition systems that improve transcription accuracy by leveraging personalized language models tailored to specific device topologies. The problem addressed is the inherent inaccuracy of generic voice recognition models when processing utterances from devices with unique configurations, such as smart speakers or other multi-microphone systems. The solution involves a two-stage transcription process. First, a voice utterance is processed using a generic language model to generate an initial transcription and a corresponding confidence score. Simultaneously, the utterance is analyzed to determine a device topology representation, which describes the spatial arrangement and characteristics of microphones or other input components. This topology data is used to select or configure a personalized language model optimized for the specific device setup. The utterance is then processed again using this personalized model to produce a second transcription and confidence score. The system compares the confidence scores of both transcriptions and selects the higher-confidence result for further interpretation. This approach enhances accuracy by adapting the language model to the unique acoustic properties of the input device, reducing errors caused by generic model assumptions. The method is particularly useful in environments where device configurations vary widely, such as smart home systems or multi-device setups.
4. The computer-implemented method of claim 1 , wherein determining, by the one or more computing devices, the one or more selected devices and the one or more corresponding actions is further based at least in part on contextual data associated with the user.
This invention relates to a computer-implemented method for selecting devices and corresponding actions in a networked environment, particularly focusing on personalizing device interactions based on contextual user data. The method addresses the challenge of efficiently determining which devices should perform specific actions in response to user inputs or system triggers, ensuring relevance and adaptability to the user's current context. The method involves analyzing contextual data associated with the user, such as location, time, preferences, or activity patterns, to refine the selection of devices and actions. This contextual data is processed alongside other factors, such as device capabilities, network conditions, and predefined rules, to dynamically determine the most appropriate devices and actions. The system may prioritize devices that are most relevant to the user's current situation, such as selecting a nearby smart speaker for audio output when the user is at home or adjusting lighting settings based on time of day. By incorporating contextual data, the method enhances user experience by ensuring that device interactions are tailored to the user's immediate needs and environment. This approach improves efficiency, reduces unnecessary device activations, and adapts to changing user contexts without manual intervention. The invention is particularly useful in smart home, IoT, or automated system applications where personalized and context-aware device control is essential.
5. The computer-implemented method of claim 4 , wherein the contextual data associated with the user comprises location data associated with the user relative to the locations of one or more of the smart devices specified in the device topology representation.
This invention relates to a computer-implemented method for managing smart devices in a networked environment, focusing on enhancing user interactions by incorporating contextual data. The method addresses the challenge of providing personalized and location-aware device control in smart home or IoT ecosystems, where users may need to interact with multiple devices based on their physical proximity or environmental context. The method involves generating a device topology representation that maps the spatial relationships between smart devices, such as their positions within a building or geographic area. Contextual data associated with the user, including location data, is collected to determine the user's position relative to the devices in the topology. This data is used to dynamically adjust device behavior, prioritize device interactions, or filter device-related information based on the user's proximity. For example, a user near a smart thermostat may receive temperature control options, while a user near a smart speaker may receive audio playback controls. The method may also involve analyzing movement patterns or historical location data to predict user intent or optimize device responses. By integrating location context with device topology, the system enables more intuitive and automated device management, reducing manual input and improving user experience in smart environments. The approach is particularly useful in multi-device scenarios where spatial awareness enhances functionality and usability.
6. The computer-implemented method of claim 1 , wherein the device topology representation further includes one or more corresponding device identifiers for each of the plurality of smart devices, and one or more corresponding device attributes for each of the plurality of smart devices.
This invention relates to computer-implemented methods for managing smart devices in a networked environment. The problem addressed is the lack of efficient and structured ways to represent and manage the topology of interconnected smart devices, including their identifiers and attributes, which is essential for automation, monitoring, and control systems. The method involves generating a device topology representation that includes a plurality of smart devices and their interconnections. The topology representation further includes one or more device identifiers for each smart device, allowing unique identification within the network. Additionally, it incorporates one or more device attributes for each smart device, such as device type, capabilities, status, or configuration parameters. These attributes enable detailed characterization and differentiation of devices, facilitating better management and interaction. The topology representation may also include hierarchical or relational data structures to depict how devices are connected or grouped, such as in a home automation system or industrial IoT network. This structured representation allows for automated discovery, configuration, and control of smart devices, improving system efficiency and scalability. The method supports dynamic updates to the topology as devices are added, removed, or reconfigured, ensuring the representation remains accurate and up-to-date. This is particularly useful in environments where device configurations change frequently, such as smart homes or industrial settings.
7. The computer-implemented method of claim 6 , wherein determining, by the one or more computing devices, the one or more selected devices and the one or more corresponding actions is further based at least in part on the one or more corresponding device identifiers specified in the device topology representation.
This invention relates to a computer-implemented method for managing device interactions in a networked system. The method addresses the challenge of efficiently determining which devices should perform specific actions based on their roles and relationships within a network topology. The system uses a device topology representation to map out how devices are interconnected and their respective identifiers. When an action is required, the method analyzes this topology to identify the relevant devices and the corresponding actions they should execute. The selection process considers the device identifiers specified in the topology representation to ensure accurate targeting. This approach improves automation and reduces manual intervention by dynamically assigning tasks based on the network structure. The method is particularly useful in environments where devices must coordinate actions in a structured and scalable manner, such as in industrial control systems, IoT networks, or distributed computing environments. By leveraging the topology representation, the system ensures that actions are delegated to the correct devices, enhancing efficiency and reliability in device management.
8. The computer-implemented method of claim 6 , wherein determining, by the one or more computing devices, the one or more selected devices and the one or more corresponding actions is further based at least in part on at least one of the one or more corresponding device attributes specified in the device topology representation.
This invention relates to a computer-implemented method for managing devices in a networked system, particularly focusing on selecting devices and determining corresponding actions based on device attributes. The method addresses the challenge of efficiently identifying and controlling devices in complex networked environments where devices have varying attributes and capabilities. The method involves analyzing a device topology representation, which includes information about the devices and their attributes, such as hardware specifications, software configurations, or operational states. Using this representation, the system determines one or more selected devices and corresponding actions to perform on them. The selection and action determination are based on the device attributes specified in the topology representation, ensuring that the chosen actions are compatible with the devices' capabilities and current states. For example, if a device has a specific attribute indicating it supports a particular communication protocol, the system may select that device and assign an action that utilizes that protocol. Similarly, if a device is in a low-power state, the system may avoid actions that would disrupt its energy-saving mode. This approach improves efficiency and reliability in device management by tailoring actions to the specific attributes of each device. The method is implemented by one or more computing devices that process the device topology representation and apply rules or algorithms to match devices with appropriate actions. This ensures that the system can dynamically adapt to changes in device attributes or network conditions, providing a robust solution for managing heterogeneous device environments.
9. The computer-implemented method of claim 1 , further comprising: determining, by the one or more computing devices, one or more control commands based at least in part on the one or more selected devices and the one or more corresponding actions, the one or more control commands respectively instructing the one or more selected devices to perform the one or more corresponding actions; and providing, by the one or more computing devices, the one or more control commands to the one or more selected devices.
This invention relates to a computer-implemented method for controlling devices in a networked environment. The method addresses the challenge of efficiently managing and executing actions across multiple interconnected devices, ensuring seamless coordination and automation of tasks. The system involves selecting one or more devices from a network and associating each selected device with a specific action to be performed. Based on these selections, the system generates control commands that instruct the selected devices to execute their corresponding actions. These control commands are then transmitted to the respective devices, enabling them to perform the desired operations autonomously. The method ensures that the devices receive precise instructions tailored to their capabilities, facilitating smooth and coordinated operation within the network. This approach enhances automation, reduces manual intervention, and improves the efficiency of device management in smart environments, industrial systems, or other interconnected device networks. The system may also include additional features such as user interfaces for device selection and action assignment, as well as feedback mechanisms to confirm successful execution of the commands.
10. A computing system, comprising: one or more processors; and one or more memory devices, the one or more memory devices storing computer-readable instructions that when executed by the one or more processors cause the one or more processors to perform operations, the operations comprising: receiving a plurality of user inputs directed at a device topology user interface used to update a device topology representation, the device topology representation defining a plurality of smart devices associated with one or more structures and a location of each of the plurality of smart devices within the one or more structures; adding, to the device topology representation and based on the plurality of inputs directed at the device topology user interface: a given smart device, a device identifier of the given smart device, the one or more structures, one or more rooms as a given location of the given smart device within the one or more structures, and one or more device attributes for the given smart device, wherein the one or more device attributes for the given smart device are associated with one or more actions the given smart device is configured to perform; subsequent to adding to the device topology representation based on the plurality of user inputs directed to the device topology representation: receiving data indicative of a voice utterance, the voice utterance provided by a user to a control device; accessing the device topology representation; determining a transcription of the voice utterance based at least in part on the device topology representation, wherein determining the transcription of the voice utterance based on the device topology representation comprises: identifying one or more biasing terms or phrases associated with the accessed device topology representation, biasing a general language model based on one or more of the identified biasing terms or phrases associated with the accessed device topology representation, and determining the transcription based on the biasing; and determining one or more selected devices, from the plurality of smart devices, and one or more corresponding actions to be performed by the one or more selected devices based at least in part on the determined transcription and the device topology representation.
The computing system is designed for managing and controlling smart devices within a structured environment, such as a home or building. The system addresses the challenge of accurately interpreting voice commands to control multiple smart devices by leveraging a detailed device topology representation. This representation includes the identity, location, and capabilities of each smart device, organized by structures (e.g., buildings) and rooms. Users interact with a user interface to update this topology, adding devices, their identifiers, locations, and attributes that define their functional actions. When a voice command is received, the system accesses the device topology to improve transcription accuracy. It identifies biasing terms or phrases from the topology (e.g., device names, room names) to adjust a general language model, ensuring the transcription aligns with the known devices and their locations. The system then determines which devices should respond to the command and the specific actions they should perform, based on the transcription and the topology data. This approach enhances voice command accuracy by contextualizing the input within the known device environment, reducing ambiguity and improving user experience.
11. The computing system of claim 10 , the operations further comprising: determining a location of the user based at least in part on a location of the control device.
A computing system is designed to enhance user interaction by leveraging the location of a control device to determine the user's position. The system includes a control device that communicates with a computing device, allowing the computing device to track the control device's location. By analyzing this location data, the system infers the user's position, enabling context-aware functionalities such as personalized content delivery, spatial navigation, or adaptive user interfaces. The system may also incorporate additional sensors or data sources to refine location accuracy. This approach addresses challenges in traditional user tracking methods, which often rely on direct user input or less precise sensors, leading to inaccuracies or inconvenience. By using the control device as a proxy for the user's location, the system provides a seamless and efficient way to deliver location-based services without requiring explicit user interaction. The technology is particularly useful in applications like virtual reality, augmented reality, smart home systems, and mobile computing, where precise user positioning is critical for optimal performance. The system may also include features to ensure privacy and security, such as data encryption or user consent mechanisms, to protect sensitive location information.
12. The computing system of claim 11 , wherein the location associated with the user is determined based at least in part on a location of the control device as specified in the device topology representation.
A computing system determines a user's location based on the position of a control device within a device topology representation. The system includes a processor and memory storing instructions that, when executed, cause the processor to generate a device topology representation of a physical environment, where the representation includes spatial relationships between multiple devices, including at least one control device. The system then determines the user's location by analyzing the control device's position within this topology. This approach enables precise user localization by leveraging the known spatial arrangement of devices in the environment. The device topology representation may be generated using sensor data, such as from cameras, microphones, or other sensors, to map device positions and orientations. The control device may be a handheld or wearable device, and its location can be used to infer the user's position, improving accuracy in applications like augmented reality, smart home automation, or industrial monitoring. The system may also track device interactions to refine location estimates over time. This method enhances user localization by integrating device topology data, reducing reliance on traditional positioning techniques like GPS or Wi-Fi triangulation, which may be less accurate indoors or in complex environments.
13. The computing system of claim 11 , wherein determining the one or more selected devices and the one or more corresponding actions is further based at least in part on the location associated with the user relative to the location of each of the plurality smart devices specified in the device topology representation.
This invention relates to computing systems for managing smart devices in a networked environment. The system addresses the challenge of efficiently selecting and controlling smart devices based on user proximity and contextual factors. The system includes a device topology representation that maps the locations and capabilities of multiple smart devices within a network. When a user interacts with the system, the computing system determines which devices to select and what actions to perform based on the user's location relative to each device's location. This ensures that only relevant devices are activated or controlled, improving efficiency and reducing unnecessary operations. The system may also consider additional factors such as device status, user preferences, or environmental conditions to refine the selection and actions. The goal is to provide a seamless and context-aware smart device management experience, where devices respond intelligently to the user's presence and needs. This approach enhances user convenience and optimizes resource usage in smart environments.
14. One or more tangible, non-transitory computer-readable media storing computer-readable instructions that when executed by one or more processors cause the one or more processors to perform operations, the operations comprising: receiving a plurality of user inputs directed at a device topology user interface used to update a device topology representation, the device topology representation defining a plurality of smart devices associated with one or more structures and a location of each of the plurality of smart devices within the one or more structures; adding, to the device topology representation and based on the plurality of inputs directed at the device topology user interface: a given smart device, a device identifier of the given smart device, the one or more structures, one or more rooms as a given location of the given smart device within the one or more structures, and one or more device attributes for the given smart device, wherein the one or more device attributes for the given smart device are associated with one or more actions the given smart device is configured to perform; subsequent to adding to the device topology representation and based on the plurality of user inputs directed to the device topology representation: receiving data indicative of a voice utterance, the voice utterance provided by a user to a control device; accessing device topology representation; determining a transcription of the voice utterance based at least in part on the device topology representation, wherein determining the transcription of the voice utterance based on the device topology representation comprises: identifying one or more biasing terms or phrases associated with the accessed device topology representation, biasing a general language model based on one or more of the identified biasing terms or phrases associated with the accessed device topology representation, and determining the transcription based on the biasing; and determining one or more selected devices, from the plurality of smart devices, and one or more corresponding actions to be performed by the one or more selected devices based at least in part on the determined transcription and the device topology representation.
This invention relates to a system for managing and controlling smart devices within a structure using voice commands. The system addresses the challenge of accurately interpreting voice commands in environments with multiple smart devices by leveraging a device topology representation that maps the location and attributes of each device. The device topology representation is updated via a user interface, allowing users to add smart devices, assign device identifiers, specify their location within structures and rooms, and define device attributes linked to actions the devices can perform. When a user provides a voice command, the system accesses this topology to bias a general language model, improving transcription accuracy by incorporating context from the device topology. The system then determines which devices and actions correspond to the transcribed command, enabling precise control of smart devices based on their location and capabilities. This approach enhances voice command accuracy by using spatial and functional context derived from the device topology.
15. The one or more tangible, non-transitory computer-readable media of claim 14 , wherein the device topology representation specifies one or more corresponding device identifiers for each of the plurality of smart devices, and one or more corresponding device attributes for each of the plurality of smart device.
This invention relates to a system for managing smart devices in a networked environment. The problem addressed is the lack of a structured way to represent and manage the topology of interconnected smart devices, including their identifiers and attributes, which is essential for efficient device discovery, configuration, and control. The system involves a computer-readable medium storing instructions that, when executed, generate a device topology representation. This representation includes a plurality of smart devices, each associated with one or more device identifiers and one or more device attributes. The identifiers uniquely distinguish each device, while the attributes describe characteristics such as device type, capabilities, or network addresses. The topology representation allows for organized mapping of devices, enabling automated discovery, configuration, and interaction within a smart network. This structured approach improves scalability and interoperability in smart device ecosystems, particularly in home automation, industrial IoT, or smart city applications. The invention ensures that devices can be dynamically added, removed, or updated while maintaining an accurate and up-to-date topology.
16. The one or more tangible, non-transitory computer-readable media of claim 15 , wherein determining the one or more selected devices and the one or more corresponding actions is further based at least in part on at least one of the one or more corresponding device identifiers specified in the device topology representation.
A system and method for managing device interactions in a networked environment involves analyzing a device topology representation to determine optimal device selections and corresponding actions. The device topology representation includes identifiers for connected devices, their relationships, and operational states. The system evaluates these identifiers to select specific devices and determine appropriate actions, such as configuration changes, data processing tasks, or communication protocols. The selection process considers device capabilities, network conditions, and predefined rules to ensure efficient and reliable operation. By leveraging the device topology representation, the system dynamically adapts to changes in the network, optimizing performance and resource utilization. This approach is particularly useful in environments with diverse device types, where manual configuration would be impractical or error-prone. The system enhances automation, reduces manual intervention, and improves overall system reliability by making data-driven decisions based on the device topology.
17. The one or more tangible, non-transitory computer-readable media of claim 15 , wherein determining the one or more selected devices and the one or more corresponding actions is further based at least in part on at least one of the one or more corresponding device attributes specified in the device topology representation.
This invention relates to a system for managing device interactions in a networked environment, particularly focusing on selecting devices and determining corresponding actions based on device attributes. The system addresses the challenge of efficiently coordinating actions across multiple devices in a network, where devices may have varying capabilities, states, or configurations. The solution involves a device topology representation that captures attributes of devices in the network, such as their capabilities, current states, or other relevant characteristics. When determining which devices to select and what actions to perform, the system considers these device attributes to ensure compatibility, efficiency, or other desired outcomes. For example, if a device lacks a required capability, the system may exclude it from selection or adjust the action accordingly. This approach enables dynamic and context-aware device management, improving system performance and reliability. The invention is implemented using computer-readable media that store instructions for executing the described processes, ensuring that the system operates in a tangible, non-transitory manner.
Unknown
June 30, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.